A glucosylcholesterol-cytoskeleton axis links GBA2 loss-of-function to synaptic and mitochondrial pathology in Hereditary Spastic Paraplegia

Background. GBA2-associated hereditary spastic paraplegia (SPG46) is a rare autosomal recessive neurodegenerative disorder caused by loss-of-function mutations in GBA2, encoding the non-lysosomal glucocerebrosidase 2. GBA2 deficiency leads to glucosylceramide (GlcCer) accumulation and glucosylated cholesterol (GlcChol) depletion, causing cytoskeletal defects in immature neurons. However, the mechanisms linking lipid dysregulation to neuronal dysfunction remain poorly understood. Methods. We modelled GBA2 loss of function by chronic pharmacological inhibition in mouse cerebellar granule neurons (CGNs) and assessed neuronal morphology, synaptic organization, Ca2+ dynamics, mitochondrial function and actin cytoskeleton during maturation. Proteomic profiling was performed in GBA2-inhibited and GlcChol-supplemented neurons. Findings were validated in a zebrafish gba2 crispant model by evaluating motor behavior, cerebellar development, neuronal organization and mitochondrial function, and in patient-derived fibroblasts carrying a homozygous pathogenic GBA2 variant (NM_020944). The role of RAC1 was studied in both neurons and patients' cultured skin fibroblasts, and upon rac1 pharmacological inhibition in zebrafish crispants. Results. Chronic GBA2 inhibition impaired axonal outgrowth in immature CGNs but not neurite complexity in mature neurons, suggesting morphological compensation. Nevertheless, mature neurons displayed enlarged presynaptic terminals, impaired synaptic vesicle clustering and altered Ca2+ responses to potassium and glutamate, the latter associated with NMDA receptor redistribution without changes in total receptor levels. Mitochondrial alterations were observed in CGNs, patient fibroblasts and zebrafish, consistent with defective architecture of the mitochondrial network. Proteomics revealed convergent alterations in actin cytoskeleton, synaptic pathways and cellular metabolism following both GBA2 inhibition and GlcChol supplementation. GlcChol bidirectionally regulated RAC1 function, likely altering its spatial distribution rather than its global activation. Confocal imaging confirmed abnormal RAC1 and F-actin localization in patient fibroblasts. Zebrafish gba2 crispants recapitulated motor deficits, Purkinje cell loss, motor neuron disorganization and mitochondrial abnormalities. Pharmacological Rac1 inhibition rescued motor behavior and neuronal organization, linking cytoskeletal disorganization to the observed phenotype in the zebrafish model. Conclusions. Our findings identify a pathogenic GlcChol-RAC1-actin signalling axis linking lipid imbalance to synaptic disorganization, NMDA receptor redistribution and mitochondrial dysfunction in SPG46. The selective vulnerability of corticospinal neurons, cerebellar granule neurons and Purkinje cells may reflect their dependence on this pathway. Rac1 inhibition rescues disease phenotypes in vivo, highlighting this pathway as a promising therapeutic target.